David G. Vaughan British Antarctic Survey Natural Environment Research Council Madingley Road Cambridge, CB3 0ET United Kingdom
Armstrong et al.(1973), defined an ice shelf in the following terms, "A floating ice sheet of considerable thickness attached to a coast. Ice shelves are usually of great horizontal extent and have a level or gently undulating surface. They are nourished by the accumulation of snow and often by the seaward extension of land glaciers. Limited areas may be aground...". This is not a very exclusive definition, and within its constraints Robin (1979) postulated that some ice shelves were derived from glaciers that had crossed their grounding lines intact and become floating, while others were derived from many years of snow accumulation on fast ice (sea ice attached to a land mass or grounded ice sheet). While the former adequately describes many large ice shelves the latter is a notable rarity, one example being the ice shelf that formerly occupied much of Prince Gustav Channel, Antarctica (Nordenskj ld 1911). A third origin for ice shelves has now been identified; Jeffries et al. (1991) showed that Ward Hunt Ice Shelf, Canada grew, at least in part, by the accumulation of sea ice beneath fast ice. It is clear that the single term ice shelf encompasses bodies of floating ice with varied origins and composition. A classification of ice shelves which would allow clear distinction to be drawn between different styles of ice shelf is proposed in this note.
Barkov (1985), presented a summary of the attempts to "systematize" our knowledge of ice shelves through discussion of various classifications which he termed morphological, morphological and kinematic, thermo-kinematic. Barkov went on to suggest his own genetic classification (sic.), based on the direction of mass transfer at the upper and lower surfaces and the type of ice comprising the ice shelf at the time of formation. This produced 16 possible categories, for which Barkov provided only a few examples. Any useful classification scheme should be based on simple, measurable parameters which indicate particular properties likely to be common among the members of each class. Below I suggest a new classification for ice shelves based on these principles.
There are three possible sources of mass input for ice shelves; glacier input, in situ surface accumulation and basal accretion. Similarly, there are three mechanisms for mass loss; iceberg calving, bottom melting and surface ablation. These dominant routes of mass input and loss give rise to six reasonable combinations which represent very different types of ice shelf, and for each of these types there are well-documented contemporary examples (Table 1). Two combinations implied by accumulation and ablation through the same surface are unlikely. There are, I believe, no good contemporary examples of the one further combination, "sustained by surface accumulation and ablated by ice berg calving" (Type B).
Although the parameters used in this classification are related simply to the dominant routes of mass transfer, there are other easily measured characteristics that should help in the diagnosis of ice shelf type (Table 2). Horizontal velocity is the most commonly measured glaciological parameter, it will be unusually low for Type E and Type G ice shelves. Surface flowlines (Crabtree and Doake, 1980) are common in visible satellite images of some ice shelves, and are commonly used to indicate the direction of flow, but they may also be an indication of type being generally absent from Type E ice shelves. Seawater saturation of an ice shelf is often inferred where radar sounding has failed to penetrate the full ice thickness (Vaughan et al., 1993), and will have a gross effect on the strength of the ice shelf. Type B and Type G ice shelves are most likely to be saturated by seawater. The predominance of either, glacier ice or firn, at the surface of an ice shelf can be inferred from satellite images (Rott, 1989) and is a good indicator of the sign of surface mass balance.
Finally, it should be noted that each of these different ice shelf types will behave differently under changing climatic conditions, with each type being likely to respond most rapidly to some particular climate variable. It is clear that Type F, for example would respond quickly to changes in ocean temperature, while Type G could also respond rapidly to changes in surface accumulation. The pattern of recent retreat of ice shelves around the Antarctic Peninsula including Wordie Ice Shelf (Type A) and Larsen A (Type A) suggests that these retreats have been in response to a rise in atmospheric temperature (Vaughan and Doake, 1996).
The largest Antarctic ice shelves, Ross, Filchner-Ronne are perhaps worthy of special mention. The Ross is most probably well classified as Type A, save that being so large it must derived considerable mass from in situ accumulation. Filchner-Ronne Ice Shelf is perhaps more complex and has considerable input from the accretion of basal ice. Indeed, we can define sections in the central part of Ronne Ice Shelf for which input is derived from glacier input, in situ surface accumulation and basal accretion, and for which there are similar rates of mass loss by basal melting and iceberg calving (Thyssen et al., 1993).
In reality, it should be noted that many ice shelves will have to be considered as hybrids between the categories described in Table 1. Many glacier-fed ice shelves, for example, derive a considerable amount of ice from in situ surface accumulation. Hell's Gate Ice Shelf, Antarctica is a hybrid with a very complex structure, described by Souchez et al. (1991). Basal accretion/iceberg calving (Type C) is present as a small fraction close to the front, but the majority of the shelf appears to follow the glacier-fed/surface melting pattern (Type D).
Despite these problems it is hoped that the classification proposed here sill serve as a reminder that all ice shelves cannot be considered as sharing the same origin or dynamical character. Models of ice sheets must take account of these differences if they attempt to assess the role of ice shelves in the past and future evolution of ice sheets.
Armstrong, T., B. Roberts and C.W.M. Swithinbank. 1973. Illustrated glossary of snow and ice. Scott Polar Research Institute, Cambridge.
Barkov, N.I. 1985. Ice shelves of Antarctica. Russian Translations Series, 20. Balkema.
Crabtree, R.D. and C.S.M. Doake, 1980. Flowlines on Antarctic Ice Shelves. Polar Rec., 20 (124), 31-37.
Doake, C.S.M., and D.G. Vaughan. 1991. Rapid disintegration of Wordie Ice Shelf in response to atmospheric warming. Nature 350 (6313): 328-330.
Jeffries, M.O., H.V. Serson, H.R. Krouse and W.M. Sackinger. 1991. Ice physical properties, structural characteristics and stratigraphy in Hobson's Choice Ice Island and implications for the growth history of East Ward Hunt Ice Shelf, Canadian High Arctic. Journal of Glaciology 37 (126): 247-260.
Nordenskj ld, O. 1911. Die Schwedische Sudpolar-expedition 1901-1903. Band1. Ihre geographische tatigkeit. Dulau and Co., London.
Potter, J.R., J.G. Paren and J. Loynes. 1984. Glaciological and oceanographic calculation s of the mass balance and oxygen isotope ratio of a melting ice shelf. Journal of Glaciology 30 (105): 161-170.
Robin, G. de Q. 1979. Formation, flow and disintegration of ice shelves. Journal of Glaciology 24 (90): 259-271.
Rott, H., 1989. Reflexionseigenschaftenvon Schnee und Eis. In: rnberger, E. Ed. Die Kartographie und ihre Randgebiete. Vol. IV/2. Wien, Deuticke, 345.
Souchez, R., M. Meneghel, J-L. Tison, R. Lorrain, D. Ronveaux, C. Baroni, A. Lozej, I. Tabacco and J. Jouzel. 1991. Ice composition evidence of marine ice transfer along the bottom of a small Antarctic ice shelf. Geophysical Research Letters 18 (5): 849-852.
Thyssen, F., A. Bombosch, H Sandh„ger, 1993. Elevation, ice thickness and structure mark maps of the central part of Filchner-Ronne Ice Shelf. Polarforschung, 62 (1), 17-26.
Vaughan, D.G., D.R. Mantripp, J. Sievers, and C.S.M. Doake. 1993. A synthesis of remote sensing data over Wilkins Ice Shelf, Antarctica. Annals of Glaciology 17: 211-218.
Vaughan, D.G. and C.S.M. Doake, 1996. Recent atmospheric warming and retreat of ice shelves on the Antarctic Peninsula. Nature 379, 328-331.